Aerial Hybrid Active-Passive Reconfigurable Intelligent Surface-Assisted Secure Communications for Integrated Satellite-Terrestrial Networks

In next-generation wireless networks, integrated satellite-terrestrial networks are regarded as a pivotal solution for supporting seamless coverage and elevated data rates, but the physical layer security performances are severely degraded under both jamming and eavesdropping attacks due to wide field of line of sight. Thus, this paper designs an aerial hybrid active-passive reconfigurable intelligent surface (aerial hybrid RIS) communication system to enhance secure and reliable communication for integrated satellite-terrestrial networks, where an active eavesdropper aims to jam legitimate channels and eavesdrop on any data stream from RIS simultaneously. Specifically, we propose a resource scheduling approach that jointly optimizes the position of the aerial RIS, the hybrid beamforming matrix, the satellite beamforming design, and the satellite transmission power to maximize the ground users’ (GUs) secrecy rate under quality of service (QoS) requirements. To address the optimization problem in complex and dynamic communication environments, we reformulate the problem as a reinforcement learning (RL) problem and propose a secure resource scheduling method based on the relay hindsight experience replay-softmax deep double deterministic policy gradients (RHER-SD3) algorithm. The proposed RHER-SD3 algorithm effectively schedules the secure hybrid active-passive beamforming matrix, the aerial position of the RIS, the satellite beamforming vectors, and the satellite power allocation to avoid both jamming and eavesdropping attacks, even though the behavior information of the attacker is imperfect. Simulation results demonstrate that the proposed method outperforms existing approaches in improving system secrecy performance and QoS satisfaction against hybrid attacks.